This project develops a new class of antimicrobial drug against capsulated bacteria. Capsules are thick layers of extracellular polysaccharides that hinder access of most antimicrobial agents to the underlying organisms. The therapeutic agents are protein enzymes - depolymerases - that degrade the bacterial capsule and expose the bacterium to attack by the immune system. These new antimicrobial agents should have moderately broad host ranges and work on bacteria resistant to classical drugs. They have the further advantage that they do not kill bacteria per se, merely enabling the immune system to do the killing, and they thus do not release endotoxins that would exacerbate symptoms. The biological sources of these enzymes used in this work are bacteriophages, viruses that kill bacteria, which are easily isolated from the wild. Instead of using intact bacteriophages, we propose to use just the enzymes, which are encoded as parts of the bacteriophage tailspike. Sporadic studies dating back 84 years suggest that the enzymes alone can effect a cure for bacterial infections. The work here will establish the generality of earlier work, developing and implementing a protocol for the isolation, identification, and calibration of depolymerase activity against arbitrary capsule types. The ultimate test provided here will be testing depolymerase efficacy in actual infections. A second direction of the work is to develop a protocol for evolving higher and broader activity of depolymerases, improving on native enzymes obtained from the wild. The work here focuses on Klebsiella pneumoniae and E. coli, with extension to Acinetobacter baumanii in the later phases.